Advanced notes on post-news reversion scalps

[LondonScalper]

This one is for people who already fade post-release spikes and want to compare notes at a finer level than "wait for the spread to settle".

I split my reversion trades by the type of release, because they behave differently. With 52 trades logged since January 2024, the numbers are still small, but the pattern is interesting.

Inflation data (CPI, PCE): the initial move tends to extend in the second and third minute before reverting, if it reverts at all. Entering on the first sign of exhaustion was a loser for me. Waiting until minute five or six improved things.

Jobs data: the first spike is often fully or largely reversed in the first two minutes on a mixed report, where the headline and the details disagree. On clean reports with everything pointing one way, reversion trades were nearly all losers.

Central bank decisions: I've stopped fading these entirely. The initial move often continues as the press conference begins.

On entry, I use limit orders at a pre-defined retracement level rather than market orders. My market order fills right after releases cost me an average of 0.6 pips more than the limit fills on comparable trades.

The other thing I track is whether the reversion happens within 15 minutes. If it doesn't, I close. Extended reversions turn into new trends more often than I expected.

[Shadow Trader]

Splitting by release type is the right level of detail. My own log is smaller than 52, but the CPI versus jobs difference matches what you saw.

On inflation I wait even longer than minute five. I want the first M1 bar that fails to make a new extreme after minute six, and I want the release price to be at least twelve pips away from that extreme on EURUSD. Under that, a third of the move does not cover costs.

Jobs is where I disagree a little. On mixed reports I will fade, but only the second spike, not the first. The first spike on NFP is often noise from the headline alone. The second one, after wages and revisions hit the wires, is the one that sometimes fails.

Central bank decisions I have also dropped for reversion. The press conference starts later and the first move is rarely the last.

Your 15 minute time stop is stricter than mine. I use 12 minutes on CPI and 8 on jobs. If it has not come back by then, I am usually late anyway.

Do you place the limit at a fixed fraction of the spike, or at a pre-marked level from before the release?

[Shadow Trader]

The abort rule I'd add is for a reversion that fails after the fill. If price prints a new extreme more than about four pips beyond the original spike inside the first three minutes after entry, I scratch immediately rather than wait for the time stop.

That came from CPI days where minute six looked exhausted, I got filled, and then a second push ran another eight pips while I sat on the 12 minute clock. The time stop was simply too slow for that shape.

So the card now has two exits. One is the usual time stop, the other is the three minute abort if a new extreme prints, and whichever hits first ends the trade.

I still only use it on inflation and mixed jobs. Clean one way jobs reports stay off the reversion list entirely, as you found.

LondonScalper, do you have an early abort when the spike extends after your fill, or only the 15 minute close? The three minute rule feels harsh until you see what the alternative costs.